Fig. 3: Introducing multiplicative noise through stochastic selector.
From: Neural sampling machine with stochastic synapse allows brain-like learning and inference
a Schematic and TEM of a fabricated stack of [Ag/TiN/HfO2/Pt] with 3 nm TiN and 4 nm HfO2. b A stochastic synapse is realized by augmenting this stochastic selector in series with the FeFET-based analog weight cell. c Measured current-voltage characteristics showing abrupt electronic transition from insulating state to metallic state due to the formation of a continuous filament of Ag+ atoms bridge the top and bottom electrodes. A wide window of variation in the threshold voltage VT that triggers the spontaneous formation of the Ag+ filament is observed. The stochasticity can be exploited by applying the input voltage Vin within the variation window of the VT. d Measured threshold voltage VT over multiple cycles. e Stochastically reading an LRS and an HRS of the FeFET through the stochastic selector. f Measured device-to-device variation across 17 selector devices. Error bar denotes standard deviation across the mean. g–i The stochasticity switching of the selector device is modeled using an Onrstein-Uhlenbeck (OU) Process. The model shows excellent agreement with the experimental data. WL word line, BL bit line, SL source line, Vin input voltage, Iout output current, LRS low resistance state, HRS high resistance state.
